Mäkinen et al. BMC Genomics (2019) 20:430 https://doi.org/10.1186/s12864-019-5817-8

RESEARCH ARTICLE Open Access Genome description of radiata 79 with comparative genomics analysis on lignocellulose decomposition machinery of phlebioid fungi Mari Mäkinen1,4, Jaana Kuuskeri1, Pia Laine2, Olli-Pekka Smolander2,5, Andriy Kovalchuk3, Zhen Zeng3, Fred O. Asiegbu3, Lars Paulin2, Petri Auvinen2 and Taina Lundell1*

Abstract Background: The white rot , a type of the Phlebia, is an efficient decomposer of plant cell wall polysaccharides, modifier of softwood and hardwood lignin, and is able to produce ethanol from various waste lignocellulose substrates. Thus, P. radiata is a promising organism for biotechnological applications aiming at sustainable utilization of plant biomass. Here we report the genome sequence of P. radiata isolate 79 originally isolated from decayed alder wood in South Finland. To better understand the evolution of wood decay mechanisms in this fungus and the phlebioid clade, gene content and clustering of genes encoding specific carbohydrate-active enzymes (CAZymes) in seven closely related fungal species was investigated. In addition, other genes encoding proteins reflecting the fungal lifestyle including peptidases, transporters, small secreted proteins and genes involved in secondary metabolism were identified in the genome assembly of P. radiata. Results: The PACBio sequenced nuclear genome of P. radiata was assembled to 93 contigs with 72X sequencing coverage and annotated, revealing a dense genome of 40.4 Mbp with approximately 14 082 predicted protein-coding genes. According to functional annotation, the genome harbors 209 glycoside hydrolase, 27 carbohydrate esterase, 8 polysaccharide lyase, and over 70 auxiliary redox enzyme-encoding genes. Comparisons with the genomes of other phlebioid fungi revealed shared and specific properties among the species with seemingly similar saprobic wood-decay lifestyles. Clustering of especially GH10 and AA9 enzyme-encoding genes according to genomic localization was discovered to be conserved among the phlebioid species. In P. radiata genome, a rich repertoire of genes involved in the production of secondary metabolites was recognized. In addition, 49 genes encoding predicted ABC proteins were identified in P. radiata genome together with 336 genes encoding peptidases, and 430 genes encoding small secreted proteins. Conclusions: ThegenomeassemblyofP. radiata contains wide array of carbohydrate polymer attacking CAZyme and oxidoreductase genes in a composition identifiable for phlebioid white rot lifestyle in wood decomposition, and may thus serve as reference for further studies. Comparative genomics also contributed to enlightening fungal decay mechanisms in conversion and cycling of recalcitrant organic carbon in the forest ecosystems. Keywords: Phlebia radiata, comparative genomics, wood decay, carbohydrate-active enzyme genes, lignin biodegradation, co-regulation, peptidases, secondary metabolism, ABC transporters, small secreted proteins

* Correspondence: [email protected] 1Department of Microbiology, Faculty of Agriculture and Forestry, Viikki Campus, University of Helsinki, FI-00014 Helsinki, Finland Full list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Mäkinen et al. BMC Genomics (2019) 20:430 Page 2 of 22

Background manganese peroxidases [23], and two laccases [24, 25] Cellulose and hemicellulose are the most abundant nat- have been cloned and characterized. Transcripts and se- ural, renewable carbohydrate polymers in plant cell walls. creted protein products of these together with an array In the biodegradation and utilization of the plant biomass, of CAZyme-encoding genes of P. radiata acting against it is considered that access to these carbohydrate polysac- plant cell wall polysaccharides were identified when the charide feedstocks is restricted due to the compact and or- fungus was grown on spruce wood [26]. dered plant cell wall lignocellulose composite including Advances in the development of next-generation genome regions of crystalline cellulose microfibrils, and the aro- sequencing platforms have enabled detailed and cost- matic and heterogenous lignin units [1]. Therefore, the use effective investigation of genomic content of diverse fungi. of plant biomass components for different biotechno- The Phlebia clade includes three genome-sequenced spe- logical applications such as production of bio-based liquid cies that are Phlebia brevispora [7], [27], fuels and chemicals requires both physico-chemical and and Phlebia radiata presented in this study. Among the enzymatic treatments to break the composite structure of Phanerochaete clade, genomes of Phanerochaete chrysos- lignocellulose [2]. porium [28], Phanerochaete carnosa [29], Bjerkandera adu- In nature, saprotrophic fungi of the Dikarya phyla Asco- sta [7]andPhlebiopsis gigantea [15] have been assembled mycota and include thousands of species re- and annotated. The number of genome-sequenced fungi sponsible for the important role in decomposition of plant representing different lifestyles has been accumulating par- litter and other biomass-based materials [3, 4]. In particular, ticularly within the 1000 Fungal Genomes project [13, 16, cycling of carbon and nutrients in the forest ecosystems is 28]. The gene content and genomic arrangement appar- largely dependent on fungal decomposition of dead wood, ently reflect the lifestyle of an organism; in fungi, compara- forest litter and soil organic compounds [5, 6]. The fungal tive genomics approaches have given explanations for e.g. phylum Basidiomycota contains species with versatile brown rot versus white rot decay of wood [8, 13, 30], and ecological roles and lifestyles such as saprotrophic, plant divergent and specific mycorrhizal associations with tree pathogenic or mutualistic mycorrhizae, and the class Agari- and other plant roots [31]. With the aid of whole-genome comycetes harnesses the most efficient decomposers of sequencing together with comparative genomics, genes en- wood lignocellulose [7, 8]. Among the ,es- coding enzymes and proteins vital for the organism in its pecially fungi in the phlebioid clade of the order Polyporales natural habitat, evolutionary relationships between fungal encompassing corticioid and polyporoid basidiocarp- species and clades may as well be predicted [8, 13, 14, 28– forming species demonstrate high potential for biotechno- 31]. Comparative genomics studies may also enable identifi- logical applications due to their ability to attack all chemical cation of novel activities important for plant biomass deg- components of lignocellulose including lignin [9, 10]. radation and deepen the overall understanding of the Phylogenetically, the Polyporales phlebioid clade fungi fungal wood decay genetics and mechanisms. may be further divided into genus level sub-clades of In this study, high-throughput genome characterization Phlebia, Phanerochaete and Byssomerulius [10, 11]. The of P. radiata as well as comparative analysis on genes most studied species with the first published Basidiomy- associated with the plant cell wall degradation among the cota sequenced genome is Phanerochaete chrysosporium phlebioid fungal species (altogether seven species ge- [12], which is considered as the model white rot fungus nomes) were elucidated. Special attention was given to the for wood decay studies. Since then, draft genomes of CAZyme-encoding genes involved in cellulose, hemicellu- several other phlebioid species, also from the genera lose and pectin degradation, and in oxidative attack on lig- Bjerkandera and Phlebiopsis, have been annotated as nin. Other genes reflecting the lifestyle of the fungus part of the DOE JGI 1000 Fungal Genomes and related including peptidase, ABC transporter, small secreted pro- community projects [13–16]. Phlebia radiata is a phle- tein (SSP) and genes involved in secondary metabolism bioid white rot fungus of the Polyporales family Merulia- were also identified in the genome and discussed. ceae [17] and the type species of genus Phlebia. The Finnish isolate P. radiata 79 has been widely studied Results showing the diverse ability to decompose and convert Characteristics of the Phlebia radiata genome assembly different wood types and plant biomass, to degrade In total, the genome assembly contained 93 contigs as harmful organic compounds, to modify lignin, and to se- unitigs at the average of 72X sequencing coverage. The crete a wide variety of carbohydrate-active enzymes assembly included the mitochondrial mtDNA [82]ina (CAZymes) and oxidoreductases [9, 18, 19]. In particu- single contig which was further removed, resulting in lar, the oxidoreductase production properties vital for final assembly of 40.4 Mbp (haploid genome size) con- biological attack on lignins and lignin-like compounds taining 92 nuclear contigs as scaffolds (Table 1). Mean have been of interest [20, 21]. Individual genes encoding read length was 5 552 bp, average contig size 436 607 class-II lignin peroxidases [22], two divergent class-II bp, maximal contig size 3 403 443 bp, minimal contig Mäkinen et al. BMC Genomics (2019) 20:430 Page 3 of 22

Table 1 Summary of the genome assembly features of the phlebioid fungal genomes included in the comparative study Genome features P. radiata P. brevispora P. centrifuga B. adusta P. gigantea P. carnosa P. chrysosporium Assembly size (Mbp) 40.41 49.96 34.85 42.73 30.14 46.29 35.15 # of contigs 92 3178 3667 1263 1195 2272 1253 # of scaffolds 92 1645 3022 508 573 1137 232 # of scaffolds ≥ 2 Kbp 92 1645 1652 508 506 1136 204 Scaffold L50 8 12 201 13 72 6 8 Scaffold N50 (Mbp) 1.98 1.31 0.05 1.03 0.12 3.53 1.91 # of gaps 0 1533 645 755 622 1135 1021 Average gene lenght (bp) 1847 1627 1378 1703 1714 1765 1684 Average transcript length (bp) 1463 1274 1134 1388 1380 1446 1401 Average exon length (bp) 227 225 274 248 230 269 259 Average intron length (bp) 72 78 80 71 69 75 66 Average protein length (bp) 488 400 378 406 411 384 408 Exons per gene 6.46 5.66 4.14 5.59 6 5.36 5.41 # of gene models 14629 16170 13785 15473 11891 13937 13602 Data were collected from the JGI MycoCosm (genome.jgi.doe.gov/programs/fungi/index.jsf) size 5 203 bp and GC content 53 %. In total, 14 082 majority of P. radiata predicted proteins were assigned to unique gene models and 547 possible splice variants the functional terms of “Catalytic activity”, “Binding”, were recognized by the aid of available RNA sequen- “Metabolic process” and “Cellular process” (Fig. 1b). cing data using the ab initio gene prediction software BRAKER1. GO enrichment analysis P. radiata genome assembly was compared with other GO enrichment analysis was conducted based on the pre- phlebioid clade fungal genomes [10, 11] including the spe- vious RNA-seq transcriptome data of P. radiata cultivated cies Phlebiopsis gigantea, Phlebia brevispora, Phlebia cen- both on solid spruce wood (2-week and 4-week time trifuga, Bjerkandera adusta, Phanerochaete chrysosporium points) and for comparison, on liquid malt extract and Phanerochaete carnosa (Table 1). Genomic data of all medium (2-week time point) [26]. GO terms over- the species are available at the Joint Genome Institute represented among the genes which were up-regulated at (JGI) MycoCosm (genome.jgi.doe.gov/programs/fungi/ both time points on spruce wood included the terms of index.jsf)[16]. Noticeable is the quality of P. radiata gen- “carbohydrate metabolic process”, “cell wall”, “oxidoreduc- ome with 0 gaps in the closed and gene annotated assem- tase activity”, “hydrolase activity, acting on glycosyl bly, and with contigs extended to <100 scaffolds (Table 1). bonds”, “extracellular region”, “external encapsulating Scaffold L50 value 1.98 Mbp is near to estimated median structure”, “catalytic activity”, “cell wall organization or average chromosomal length (between 0.8 – 3.5 Mbp). biogenesis” and “lipid metabolic process” (Additional file The species P. gigantea, P. centrifuga and P. chrysospor- 1). The enriched term of “carbohydrate metabolic process” ium had somewhat smaller genome sizes (30-35 Mbp) was statistically the most significant (P-value 1.03 x 10-19). than the other phlebioid species. The average length of P. At the four week time point on spruce wood substrate, radiata gene and transcript models were the largest the most enriched GO terms among the up-regulated among the studied fungi (1847 bp and 1463 bp, respect- genes included the “transmembrane transport” and “estab- ively). Noticeable is also the highest number of recognized lishment of protein localization to membrane”.Much intron-exon junctions in the gene models of P. radiata greater number of GO terms were enriched only among (average exon number 6.46 per gene; Table 1). the genes which were up-regulated at the 2-week time point on spruce wood. These included for example GO Functional annotation of the P. radiata gene models terms referring to organic acid metabolism (“carbox- The presence of 5´ upstream open reading frames (uORFs), ylic acid metabolic process”, “organic acid metabolic high number of introns and very short exons (from a few process”, “oxoacid metabolic process”) and to transla- to even less than one codon in some cases) were found to tion (“structural constituent of ribosome”, “ribosome”, be typical for P. radiata genes. For translated 365 gene “translation”, “ribonucleoprotein complex biogenesis”, models, no blastp or protein domain hits were obtained. “intracellular ribonucleoprotein complex”, “ribosome According to Gene Ontology classifications (Fig. 1a, b), the biogenesis”, “ribonucleoprotein complex”). Mäkinen et al. BMC Genomics (2019) 20:430 Page 4 of 22

Fig. 1 Gene Ontology classification of P. radiata proteins. Functional classification by GO categories (a) and GO terms (b)

Carbohydrate-active enzyme genes In total, 209 glycoside hydrolase (GH), 27 carbohydrate Carbohydrate-active enzyme encoding genes of P. esterase (CE) and 8 polysaccharide lyase (PL) genes were radiata were annotated previously during the combin- identified in the P. radiata genome (Additional file 2). Of atory transcriptome and proteome study on wood [26]. these genes, 76 GHs, 15 CEs and 1 PL were annotated to Consequently, a comparative analysis on the number of putatively encode enzymes involved in breakdown of plant CAZyme-encoding genes of P. radiata and six other se- cell wall lignocellulose. In addition, twelve auxiliary activ- quenced phlebioid species of Polyporales fungi (Table 2) ity family 9 (AA9) lytic polysaccharide monooxygenase was performed. (LPMO) encoding genes and three genes coding for AA14 Mäkinen et al. BMC Genomics (2019) 20:430 Page 5 of 22

Table 2 Number of the identified CAZyme-encoding genes involved in breakdown of plant cell wall lignocellulose components in the genomes of the studied phlebioid fungi Cellulose active P. radiata P. brevispora P. centrifuga B. adusta P. gigantea P. carnosa P. chrysosporium GH1 2 2 2 2 2 2 2 GH3 10 8 7 9 9 11 10 GH5_5 4 4 2 4 4 6 2 GH5_22 2 2 2 2 2 2 2 GH6 1 1 1 1 1 1 1 GH7 6 4 3 5 5 5 8 GH9 1 1 1 1 1 1 1 GH12 2 2 2 2 3 3 2 GH44 1 2 1 0 0 0 0 GH45 3 3 1 1 1 1 2 GH131 2 4 2 3 2 2 3 Hemicellulose active CE1 2 1 1 1 2 3 4 CE5 0 1 0 0 0 0 0 CE15 2 2 6 2 1 3 2 CE16 10 8 6 15 6 5 7 GH2 3 2 3 3 3 2 2 GH5_7 3 2 3 2 2 2 3 GH10 7 8 6 4 4 5 6 GH11 1 0 0 0 2 1 1 GH27 2 2 2 3 3 3 3 GH29 1 1 1 0 0 0 0 GH31 4 5 6 4 6 8 6 GH35 3 4 3 4 2 4 3 GH43 2 2 2 6 7 4 4 GH51 1 1 4 2 2 2 2 GH74 2 1 1 2 2 2 4 GH95 1 3 1 1 0 1 1 GH115 1 2 2 2 1 1 1 Pectin active CE8 2 3 1 2 4 2 2 CE12 0 0 0 1 1 0 0 GH28 7 5 5 6 10 4 5 GH53 1 1 1 1 1 1 1 GH78 2 1 2 2 1 1 1 GH88 1 1 1 1 1 1 1 GH105 3 0 2 1 0 0 0 PL1 0 0 0 1 0 0 0 PL4 1 1 1 1 0 0 0 Auxiliary redox enzymes AA1 7 11 6 2 5 10 5 AA2 10 15 8 21 9 11 15 AA3 32 39 23 39 23 36 38 Mäkinen et al. BMC Genomics (2019) 20:430 Page 6 of 22

Table 2 Number of the identified CAZyme-encoding genes involved in breakdown of plant cell wall lignocellulose components in the genomes of the studied phlebioid fungi (Continued) Cellulose active P. radiata P. brevispora P. centrifuga B. adusta P. gigantea P. carnosa P. chrysosporium AA3_1 (CDH) 1 1 1 1 1 1 1 AA5 10 8 6 7 6 6 7 AA9 12 12 11 28 15 11 16 AA14 3 2 2 3 3 2 2 DyP 1 3 2 10 4 1 0 HTP 5 2 4 4 4 3 3 CAZy GH, CE, PL and AA classes, dye-decolorizing peroxidases (DyP), and heme-thiolate peroxidases (HTP) are indicated lytic xylan oxidases were recognized in the genome. The oxidoreductase family AA9 (Fig. 2). Typical for Polyporales carbohydrate-binding module (CBM) for cellulose- white rot fungi, one gene encoding GH6 (non-reducing binding domain (IPR000254) was detected in 32 genes end active cellobiohydrolase) was conserved in all phlebioid coding for AA9 LPMOs, and exoglucanases of the families genomes whereas the number of genes coding for GH7 GH6, GH7 and GH131, one GH3 β-glucosidase, endoglu- (reducing end active cellobiohydrolases) ranged from three canases of the families GH5 and GH45, xylanases of the up to eight (Table 2). P. radiata genome harbors six genes families GH10 and GH11, GH5 mannanases, one CE1 coding for GH7 proteins, of which three are unique and acetyl xylan esterase, one CE15 glucuronoyl esterase, and three have shared protein homology, indicating sequential one CE16 acetylesterase (Additional file 2). In addition, four gene duplications. The gene models minus.g2003 and CBM1 domain-containing proteins with no homology to plus.g2026 contain only 10 and 14 nt (nucleotide) differ- CAZyme activity catalytic domains were detected in P. ences, respectively. The third gene model minus.g8589 radiata. One of these genes encodes a putative contains two 44 nt differences and 2 nt deletions resulting carbohydrate-binding iron reductase whereas another pro- in e.g. variation in translation for two codons in compari- tein with CBM1 demonstrated homology to son to the other two (together identical) protein models. phosphodiesterases. The single nucleotide polymorphism (SNP) detected among the duplicates occurs at the codon third position. Cellulose breakdown Variations in the GH7 proteins of P. radiata and the other As could be expected, all the studied phlebioid genomes phlebioid fungi are frequent in the protein model linker re- possess genes encoding activities for the hydrolytic decom- gion separating the catalytic domain and fused CBM1 position of polymeric cellulose including cellobiohydro- domain. lases, endoglucanases and β-glucosidases (Fig. 2). Largest CAZy family GH5 comprises hydrolytic activities on number of cellulolytic genes was found in the CAZy fam- various polysaccharide substrates present in the plant ilies GH3 and GH7 as well as in the auxiliary cell wall and in the fungal cell wall [32]. Hence, GH5

Fig. 2 Number of cellulose decomposing-enzyme coding genes in the genomes of the phlebioid fungi Mäkinen et al. BMC Genomics (2019) 20:430 Page 7 of 22

sub families 5, 7 and 22 were included in this compara- Hemicellulose degradation tive study. Family 5_5 includes endo-β-1,4-glucanase Genomes of the phlebioid fungi possess a wide array of (EC 3.2.1.4) activities, family 5_7 endo-β-1,4-mannanase genes encoding all the necessary activities for the decom- activities (EC 3.2.1.78) and family 5_22 activities of endo- position of glycosidic and ester linkages of diverse hemicel- beta-1,4-glucanase and β-xylosidase. Of the seven phle- luloses (Fig. 3). Largest number of genes were detectable in bioid genomes, P. centrifuga and P. chrysosporium possess the carbohydrate esterase family CE16 (acetylesterase activ- only two GH5_5 endo-β-1,4-glucanase encoding genes ity) ranging from 5 to 15 genes per genome (Table 2), and whereas the other genomes contain at least four genes glycoside hydrolase families GH10, GH31 and GH43 (Fig. (Table 2). Interestingly, the family GH44 endoglucanases 3). However, the GH29 α-L-fucosidase encoding genes ap- were identified as one or two gene copies only in the three pear to be specific for the three Phlebia species (Table 2) Phlebia species. P. radiata and P. brevispora have three whereas the GH11 endoxylanase encoding genes were ab- genes encoding GH45 endoglucanase, whereas the other sent in the genomes of P. brevispora, P. centrifuga and B. phlebioid fungi contained less, one or two genes. Note- adusta. Interestingly, P. gigantea genome harbors two worthy is that all phlebioid genomes possess one copy of a genes for GH11 endoxylanase while missing the GH95 α- putative GH9 cellulase encoding gene. fucosidase activity genes. A putative CE5 acetyl xylan ester- All phlebioid fungal genomes apparently harbor two ase gene was detected only for P. brevispora. GH1 β-glucosidase-encoding genes, whereas seven or Genes encoding the novel AA14 oxidoreductase family of more GH3 family β-glucosidase-encoding genes were lytic xylan oxidases/monooxygenases were identified in P. identified (eight genes in P. radiata)(Table2, Fig. 2). brevispora, P. gigantea, B. adusta, P. chrysosporium and P. CAZy family GH3 is a complex family including also carnosa in a previous study [35]. For P. radiata,threeputa- activities affecting the fungal cell wall [33]. Two to tive genes with translated amino acid sequence homology three GH12 family genes encoding putative endoglu- to the Pycnoporus coccineus AA14 proteins were identified canases and xyloglucan-specific endo- β-1,4-gluca- by blastp search of the genome protein models (Table 2). nases were equally present in the phlebioid fungi. In For P. centrifuga, two possible AA14 homologs were found addition to the traditional endoglucanase and by blastp search against the MycoCosm genome data. Inter- cellobiohydrolase-encoding genes, the phlebioid ge- estingly, two of the P. radiata AA14 genes encode proteins nomes contain 2-4 putative members of the family with an extension in the C-terminus (3´ end) whereas the GH131 broad specificity β-glucanases. One of the two third gene protein model is shorter in the C-terminus than P. radiata GH131 protein models included a C- the corresponding P. coccineus AA14 homologs. All terminal CBM1 domain indicating possible attach- three P. radiata AA14 protein models demonstrate a ment to polymeric cellulose. In addition to the hydro- conserved N-terminal histidine after the signal se- lytic cellulolytic enzymes, all phlebioid genomes quence cleavage site, similar to the features of the P. harbor a large number (11-28) of AA9 lytic polysac- coccineus proteins [35]. Whereas one of the AA14 charide monooxygenase encoding genes (Fig. 2)in- genes was highly up-regulated in P. coccineus culti- volved in oxidative decomposition of cellulose [34]. vated on pine and aspen substrates [36], the P.

Fig. 3 Number of genes encoding hemicellulose-decomposing enzymes in the genomes of the phlebioid fungi Mäkinen et al. BMC Genomics (2019) 20:430 Page 8 of 22

radiata homologs were more constitutively expressed II versatile peroxidase (VP) [13, 14]. Auxiliary enzymes on spruce wood [26] (Additional file 2). possibly involved in providing hydrogen peroxide for the lignin-modifying peroxidases include, for example, family Pectin degradation AA3 glucose-methanol-choline (GMC) family oxidases From the CAZy families including pectin-decomposing ac- (glucose/aryl alcohol/alcohol/pyranose oxidases) and AA5 tivities, GH28, GH105 and CE8 families had the largest copper radical oxidases (galactose/glyoxal oxidases). number of genes (Fig. 4).ThenumberoffamilyGH28 Members of these gene families were present in the phle- polygalacturonase and rhamnosidase encoding genes was bioid genomes. Importantly, one copy of the cellulolytic relatively large (4-10 genes) in the phlebioid genomes, and lignin-attacking oxidative activity interlinking AA3_1 especially in the P. gigantea genome (Table 2). Also, four oxidoreductase family cellobiose dehydrogenase (CDH) genes encoding CE8 pectin methylesterases were identified encoding gene is present in all of the phlebioid genomes. in P. gigantea genome. Differences were noticed in the In addition, P. radiata genome includes one gene encod- CE12 family including acetylesterase activity against pectin ing a putative dye-decolorizing peroxidase (DyP) and five with genes detected only in B. adusta and P. gigantea,and genes encoding putative chloroperoxidase-like heme- family GH105 rhamnogalacturonyl and glucuronyl hydro- thiolate peroxidases (HTP) [37]. Presence of a few putative lase activity encoding genes which were identified (1-3 HTPs and 1-4 DyPs is a common feature among the phle- genes) only in the genomes of P. radiata, P. centrifuga and bioid fungi (Table 2), thereby extending their functional B. adusta. Of the potential pectin-acting lyases, one gene potential in oxidation and activation of lignin-like and encoding family PL4 rhamnogalacturonan endolyase activ- various phenolic compounds [37]. Exceptional is B. adusta ity was restricted to the three Phlebia species and B. adu- with 10 potential DyP encoding genes recognized in the sta, whereas family PL1 pectin/pectate lyase (one gene) genome, and the absence of DyP genes in P. chrysospor- was specific for B. adusta (Fig. 4,Table2). ium (Table 2,Fig.5).

Lignin conversion Genomic clustering of CAZyme genes In total, 65 CAZy auxiliary activity (AA) class genes en- Carbohydrate-active enzyme encoding genes under con- coding lignin-modifying and associated redox activities sorted expression and regulation mechanisms were were identified in P. radiata genome (Additional file 2). searched from the P. radiata genome assembly accord- All seven phlebioid species harness multiple AA1 lac- ing to the criteria of (i) close genomic coordinates, and cases/ferroxidases/laccase-like multicopper oxidases and (ii) up-regulation of gene expression on spruce wood as multiple AA2 family class-II heme-including, lignin and growth substrate in our pervious study [26] (Additional manganese peroxidases (LiP, MnP) (Table 2, Fig. 5). Not- file 3). The most interesting clusters include genes en- ably, the three Phlebia species include several genes en- coding activities against plant cell wall components and coding both long-MnP and short-MnP AA2 peroxidases genes showing strong co-regulation. P. radiata CAZyme [14, 23] together with lignin peroxidase encoding genes, genes were identified and annotated in 33 unitigs whereas B. adusta genome encodes one additional class- whereas the majority (54 %) of the CAZyme genes were

Fig. 4 Number of pectin-decomposing enzyme coding genes in the genomes of the phlebioid fungi Mäkinen et al. BMC Genomics (2019) 20:430 Page 9 of 22

Fig. 5 Number of lignin-modifying and auxiliary oxidoreductase-coding genes in the genomes of the phlebioid fungi assigned to seven distinct unitigs. When applying the 83, three AA9 genes and a non-CAZyme protein encod- sliding window approach [38], 55 CAZyme genes were ing gene (minus.g10275) form a cluster (Fig. 6c). From assigned to clusters. Minimum of three CAZyme genes the three AA9 genes, two are strongly up-regulated on in a window of maximum of 11 genes was considered as spruce whereas the third AA9 gene was up-regulated a CAZyme cluster. The size of the cluster window was only at the two-week time point of spruce cultivation. In calculated by dividing the number of P. radiata gene unitig 6, another set of additional two AA9 genes are lo- models (14082) with the number of annotated CAZyme cated next to each other but only one of the genes was genes (244). Hence, when random distribution of up-regulated on spruce wood (Additional file 3). CAZyme genes was expected, one CAZyme gene should Clustering of GH10 enzyme encoding genes was simi- appear once within 58 subsequently ordered genes. larly recognized to be shared by the three species of the When the three CAZyme genes of the cluster are genus Phlebia (P. radiata, P. brevispora, P. centrifuga)but expected to be separated from each other by less than a not in the three other phlebioid fungi studied (Table 3). fifth of the average distribution number, the size of the Clustering of 2-6 GH10 genes is interrupted in the Phlebia window was 11 (0.2 x 58). A total of 17 clusters were genomes by a maximum of three non-CAZyme protein detected according to these criteria (Additional file 3). In encoding genes. However, considering AA9 LPMO encod- addition, interesting clusters including two CAZyme ing genes, majority of the phlebioid species demonstrated genes next to each other and encoding plant cell wall several AA9 gene clusters in different scaffolds (2-4 clus- degrading activities are also included in the data set ters) except for the P. centrifuga showing only one cluster (Additional file 3). of two AA9 encoding genes (Table 4). In general, no clear tendency for tight clustering Other CAZyme gene clusters of interest and identified according to genomic localization could be asserted to in P. radiata genome included, for example, two GH45 highly similar expression patterns of the multiple genes endoglucanase encoding genes located next to each encoding diverse CAZy classes and families. However, other in unitig 0 while only one of the genes was clustering, and in some cases, concerted expression and strongly up-regulated on spruce wood (Additional file 3) possible co-regulation of the members of GH10 (xyla- . Of three CE16 carbohydrate esterase encoding genes nase activity) and AA9 (LPMO) family genes was de- forming a cluster in the same unitig, two were up- tected (Figure 6). There is a cluster of four putative regulated on spruce wood whereas one gene was down- GH10 endo-beta-1,4-xylanase encoding genes in near lo- regulated. Two up-regulated GH35 genes were identified cation and antisense orientation in the unitig 9 of P. next to each other in unitig 23. In addition, two GH3 β- radiata genome assembly (Fig. 6a). Three of the genes glucosidase encoding genes located next to each other in were up-regulated on spruce wood substrate. In unitig 9, unitig 59 and two GH28 encoding genes next to each two AA9 lytic polysaccharide monooxygenase encoding other in unitig 96, but these genes were not co-regulated genes are neighbours and strongly up-regulated at both (one gene was up-regulated while the other was down- time points on spruce wood substrate (Fig. 6b). In unitig regulated) thereby indicating individual regulation of Mäkinen et al. BMC Genomics (2019) 20:430 Page 10 of 22

Fig. 6 Clustering and co-regulation of GH10 endo-1,4-β-xylanase (a) and AA9 LPMO encoding genes (b, c)ofP. radiata. Intensity of the color indicates the scale of up- or down-regulation (dark red for high up-regulation, green for down-regulation, yellow for no significant difference) of gene expression on spruce wood substrate [26]. Direction of the arrow indicates gene location and orientation (on plus or minus strand) in the unitigs of the genome assembly (this study). The arrowhead indicates the 2-week time point of spruce wood cultivation whereas the end of the arrow refers to the 4-week time point. Beginning and end of the cluster (as genomic coordinates) is marked above the unitig. Gene model minus.g10275 in unitig 83 encodes a non-CAZyme protein their expression. On the contrary, two family GH5 anno- regulated on spruce wood substrate [26]. Similarly, the tated genes were found adjacently located in unitig 117 AA1 laccase encoding lacc1 and lacc2 genes of P. radiata and convergently expressed - up-regulated at both time located in unitig 55 are neighbours and possess similar ex- points on the spruce wood substrate. Moreover, both of pression profiles but, however, were not strongly up- the encoded proteins demonstrated high homology with regulated on spruce wood [26]. In unitig 70, one AA8 iron a specific P. chrysosporium β-xylosidase that cleaves reductase encoding gene is seemingly co-expressed with a xylan hemicellulose chains synergistically with endo- GH128 β-glucanase encoding gene. Interestingly, the only xylanases [39]. cellobiose dehydrogenase (CDH) encoding gene of P. Of the ten identified AA2 class-II peroxidase encoding radiata was recognized to situate next to a putative cata- genes of P. radiata targeted to lignin modification (Table lase encoding gene in the same unitig 70, with concurrent 2), lip3 and lip4 genes are closely located in the genome, up-regulation at both time points on spruce wood. Of the and separated only by one gene encoding a hypothetical potential hydrogen peroxide producing extracellular en- protein in unitiq 9. Surprisingly, the genes were not co- zymes, three AA5 copper radical oxidase encoding genes

Table 3 Genomic clusters of GH10 endoxylanase encoding genes in the genome assemblies of the phlebioid fungi Fungus Cluster location in the genome assembly Number of genes in the cluster Number of non-CAZyme genes in the cluster P. radiata Unitig 9/Scaffold 8:672573-681272 4 0 P. brevispora Scaffold 12:834597-856661 9 3 P. centrifuga Scaffold 2925:6494-13943 3 0 B. adusta absent - - P. gigantea absent - - P. carnosa Scaffold 3:3347462-3351815 3 1 P. chrysosporium absent - - Locations of P. radiata gene models are indicated as original assembly unitig positions, and as scaffold positions of the MycoCosm genome repository. Other data were retrieved from the MycoCosm Mäkinen et al. BMC Genomics (2019) 20:430 Page 11 of 22

Table 4 Genomic clusters of AA9 encoding genes in the genome assemblies of the seven phlebioid fungi Fungus Cluster location in the genome assembly Number of genes in the cluster Number of non-CAZyme genes in the cluster P. radiata Unitig 6/Scaffold 23:237827-242972 2 0 Unitig 9/Scaffold 8:1780628-1784094 2 0 Unitig 83/Scaffold 30:124277-133397 4 1 P. brevispora Scaffold 7:1939036-1942768 2 0 Scaffold 8:1713073-1722627 4 1 Scaffold 9:703327-712085 3 1 P. centrifuga Scaffold 809:16725-20452 2 0 B. adusta Scaffold 3:699996-705149 3 0 Scaffold 12:565705-571395 3 0 Scaffold 19:267134-284710 9 2 Scaffold 35:28836-31774 2 0 P. gigantea Scaffold 13:72697-82083 4 2 Scaffold 167:19857-23789 2 0 Scaffold 240:27459-32596 3 0 P. carnosa Scaffold 3:465671-473281 3 0 Scaffold 10:958421-970775 4 2 P. chrysosporium Scaffold 7:1587181-1592893 4 1 Scaffold 10:522449-536923 7 4 Locations of P. radiata gene models are indicated as original assembly unitig positions, and as scaffold positions of the MycoCosm genome repository. Other data were retrieved from the MycoCosm were identified in unitig 87 as separated by two interven- domains. One of the 49 genes is apparently disrupted by ing non-CAZyme genes. Two of these AA5 genes were a retrotransposon insertion (ABCF3 encoding gene) up-regulated on spruce, whereas the third gene in the (Additional file 7). All types of ABC transporter encod- middle of the cluster was down-regulated. The above de- ing genes, which are commonly found in the species of scribed clustering of functionally related or similar genes Basidiomycota class Agaricomycetes could be identified suggest that despite their close structural location in the [41, 42]. Not surprisingly, the set of putative ABC pro- genome, genes in the same cluster may display individual teins of P. radiata demonstrates high similarity with the and deviated pattern of regulation of expression. set of ABC proteins found in the genomes of P. brevis- pora [41] and P. centrifuga (this study). The differences Peptidases between the three species of Phlebia are merely quanti- Annotation of the putative P. radiata peptidases according tative (Table 6), probably due to a few lineage-specific to the classifications in the MEROPS database [40] cases of gene loss or duplication events. (https://www.ebi.ac.uk/merops/) indicated that at least 327 Compared with the four other phlebioid clade species (B. genes identified in the P. radiata genome possibly encode adusta, P. gigantea, P. carnosa, P. chrysosporium), a few proteolytic enzymes (Table 5). Highest number of the genes notable differences may be noticed (Additional file 8). The were classified to code for serine peptidases. Among the three species of Phlebia all possess one gene encoding a seven phlebioid species, P. radiata was the only species of transporter of ABC-A family, whereas corresponding genes the genus Phebia comprising genes (2 genes) encoding are missing from the other phlebioid species (Table 6). putative glutamic peptidases. However, more than ten ABC-A transporters are likely involved in the transport of genes encoding glutamic peptidases were detected in the lipid molecules, and they were previously identified in most two species of Phanerochaete,andinP. gigantea. of the analyzed fungi of the Basidiomycota subphylum Agaricomycotina [41, 42]. Likewise, a soluble ABC protein ABC transporters classified as ABCF3 encoding gene was identified in the ge- In total, a versatile set of 49 genes encoding predicted nomes of P. brevispora, P. centrifuga and P. radiata (Table ABC proteins were identified in the genome assembly of 6, Additional file 8). The function of ABCF3 encoding gene P. radiata (Table 6, Additional file 7). Out of them, 40 is unknown, and it is apparently non-essential, since the genes encode predicted ABC transporters, and 9 genes gene is absent from ca. 40% of the analyzed species of encode ABC proteins without any transmembrane Agaricomycotina. We could not identify the gene in the Mäkinen et al. BMC Genomics (2019) 20:430 Page 12 of 22

Table 5 Number of peptidase encoding genes identified in the genomes of the seven phlebioid fungi Peptidase family Fungal species Aspartic Cysteine Glutamic Metallo Serine Threonine Sum of peptidases P. radiata 48 50 2 60 149 18 327 P. brevispora 67 59 0 73 207 19 425 P. centrifuga 58 45 0 51 138 18 310 B. adusta 60 69 0 68 173 18 388 P. gigantea 43 51 15 62 159 18 348 P. carnosa 70 91 34 61 170 17 443 P. chrysosporium 64 45 11 55 149 19 343 Data was searched and retrieved from the MycoCosm

Phanerochaete clade, and in P. radiata, the corresponding and Pks5), whereas two genes are predicted to encode gene is disrupted by an insertion of a retrotransposon, a type III PKS (Pks6) and a hybrid PKS-NRPS (non- thereby most likely making it non-functional. Another gene ribosomal peptide synthase) (Pks3), respectively. Simi- classified as ABCG6.2 and encoding an ABC transporter of lar hybrid PKS-NRPS-encoding genes are found in a unknown function was found in P. chrysosporium, P. car- number of wood-inhabiting Basidiomycota Agaricomy- nosa and P. gigantea, but not in the species of Phlebia.This cetes fungi including Heterobasidion irregulare [43]. gene, however, was previously identified in less than 25% of Based on the presence of conserved modules recog- the analyzed species of Basidiomycota [41, 42]. Another nized in the deduced proteins, the identified type I notable feature of the set of ABC genes in P. radiata is a PKSs could be further classified into one non- high number of ABCC4 encoding genes, present in eight reducing PKS (Pks2) and three reducing PKSs. copies in the most expanded subfamily of ABC-C proteins Among the three reducing type I PKS genes, Pks4 (Table 6, Additional file 7). and Pks5 share 73 % of amino-acid sequence identity, suggestive of a gene duplication. Two of the identified Secondary metabolism genes of P. radiata PKS encoding genes form larger SM pathway gene Bioinformatic analysis of secondary metabolite (SM) clusters. The non-reducing type I PKS-encoding Pks2 biosynthesis-related genes of P. radiata indicates that gene was identified next to a NRPS-like encoding the genome harbors a rich repertoire of genes puta- Nrl6 gene (see below for the discussion of this gene tively involved in the production of secondary metab- family) together with three genes encoding cyto- olites. At this stage, however, no accurate chemical chrome P450 redox enzymes and one predicted gene structures of the potential products could be inferred coding for a glycosyl transferase. Interestingly, GH76 (Additional file 4). and GH16 CAZyme genes are located close to Pks2 gene. Pks4 was observed to be clustered with the Polyketide synthases NRPS-like gene Nrl8, in addition to two genes encod- Six genes encoding predicted polyketide synthases ing MFS transporters and several genes encoding pu- (PKSs) were identified in the genome assembly of P. tative tailoring enzymes, indicative of functional roles radiata (Pks1-6) (Additional file 4). Four were anno- in chemical modifications of the compounds produced tated as type I PKS encoding genes (Pks1, Pks2, Pks4 by the corresponding PKS enzymes.

Table 6 Number of ABC transporter encoding genes identified in the genomes of the seven phlebioid fungi Fungal species / ABC Subfamily P. radiata P. brevispora P. centrifuga B. adusta P. gigantea P. carnosa P. chrysosporium ABC-A 1 1 1 0 0 0 0 ABC-B (FL) 3 3 4 4 3 3 3 ABC-B (HT) 8 6 6 8 9 7 8 ABC-C 18 26 20 13 19 18 18 ABC-D 2 2 2 2 2 2 2 ABC-E 1 1 1 1 1 1 1 ABC-F 5 5 5 4 5 5 5 ABC-G 8 8 7 7 6 8 10 ABC-I 3 3 3 3 3 3 3 Data was searched and retrieved from the MycoCosm Mäkinen et al. BMC Genomics (2019) 20:430 Page 13 of 22

Adenylate-forming reductases enclose a NRPS-specific C-terminal domain. The func- Members of this protein family resemble non-ribosomal tion of these proteins is still obscure. peptide synthases (NRPS), but lack a condensation do- Among the multiple predicted proteins involved in main [44]. Although no canonical NRPS genes were biosynthesis of secondary metabolites in P. radiata, six identified in the genome of P. radiata, however, nine genes genes coding for Nrl4, Pks6, Ter1, Ter3, Ter4 and Ter12 encoding predicted adenylate-forming reductases were were earlier recognized as being up-regulated in the P. found. Based on their domain organization, one gene (Lys2 radiata transcriptomes on spruce wood substrate (at 2- homolog) was identified as a putative L-α-aminoadipate re- week and 4-week time points) [26] (Additional file 4). ductase, which is involved in L-lysine biosynthesis in fungi This indicates specific functions and involvement of [44]. The remaining eight genes (Nrl1-8) encode adenylate- chemically diverse fungal secondary metabolite com- forming reductases of unknown function, and several of pounds in active hyphal growth and colonization of them are clustered with other SM encoding genes: Nrl1 is wood, and potential participation in the wood decom- clustered with a predicted terpene cyclase, Nrl6 and Nrl8 position processes. are located close to the predicted PKS genes, Nrl7 is adja- cent to a cytochrome P450 encoding gene, whereas the Small secreted proteins of P. radiata genes for Nrl2, Nrl3 and Nrl4 form a cluster on their own, A total number of 956 genes identified in the P. radiata with near location to the gene for Nrl5. Interestingly, a genome were predicted to encode secreted proteins, among GH12 endoglucanase gene, putative AA3 alcohol oxidase which 215 represented small and cysteine-rich proteins and aryl-alcohol oxidase encoding genes, and a GH131 cel- (small secreted proteins, SSPs). From the in silico predicted lulase gene are found close to the cluster of Nrl2, Nrl3 and protein secretome, 83 and 307 proteins were selected by Nrl4. In addition, the gene for Nrl7 adenylate-forming re- EffectorP and LOCALIZER, respectively, as putative SSP ductase was recognized to be located next to an AA9 fam- candidates. Afterwards, all SSP candidates were merged ily LPMO encoding gene. Interestingly, expression of this andtheannotatedandpredictedCAZymeswereremoved AA9 gene was not up-regulated on spruce wood substrate from the final analysis, thereby resulting in a final set of [26] indicating a distinct function for the specific AA9 430 candidate proteins as SSPs (Additional file 5). enzyme other than attack on the cellulose polymer. Within The Venn diagram depicts the number of SSPs obtained the gene clusters involved in SM biosynthesis, genes according to the three criteria used (Fig. 7). Eight SSPs encoding various oxidases indicate their functional activity (genes minus.g2681, minus.g6180, plus.g1485, plus.g2909, in modifications of the products of the SM pathways. plus.g2957, plus.g7370, plus.g8107, plus.g881) were shared by all criteria, and could be deemed as good SSP candi- Terpene cyclases dates for functional validation after gene structural inspec- Predicted terpene compound forming terpenoid cyclases tion (Table 7). In the set of predicted SSPs in P. radiata, (TC) [45] constitute the most numerous group of the the shortest SSP candidate was only 66 amino-acids in identified SM biosynthesis involved genes in the genome length (gene plus.g1238.t1), while the longest protein was of P. radiata. In total, 16 putative TC encoding genes 1718 aa in length (gene plus.g13035.t1). The average (Ter1-Ter16) were identified, of which one (Ter16) could length for the translated proteome, secretome and the se- encode a squalene synthase. The products of remaining lected SSPs was in descending order (487, 394 and 353 aa, genes are unknown. Many of the TC genes are located in Additional file 6), opposite to the average of cysteine com- clusters: Ter5, Ter6 and Ter7 and Ter8 – Ter11 form two position in the protein models, which were 1.49 %, 2.12 % separate clusters on unitig70; Ter14 clusters with Ter15, and 3.01 % for the three protein sets, respectively. Accord- and Ter1 is located next to the adenylate-forming reduc- ing to the P. radiata transcriptome data [26] (Additional tase encoding gene Nrl1. There is a CAZyme gene cluster file 5), the top up-regulated SSP genes were found to en- close to the cluster of Ter8-Ter11 including two GH5 code putative hydrophobins. Hydrophobins are surface- genes and a putative AA3 family alcohol oxidase encoding active small secreted proteins that have diverse roles in gene. Ter2 gene is also located between two AA3 family fungal hyphal and mycelial growth and lifestyles [46]. genes and next to a cytochrome P450 gene. Discussion AMP-dependent acyl-CoA synthetases Basidiomycota species of the phlebioid clade of the system- Our bioinformatics analysis identified two additional atic order Polyporales [10, 11] are wood-inhabiting fungi genes coding for unknown functions, which, based on with biotechnological interest due to their ability to decom- their predicted protein domain organization, could have pose all components of wood lignocelluloses, including a role in secondary metabolism. Both genes share some modification and degradation of wood lignin [5, 12, 26]. similarity with AMP-dependent acyl-CoA synthetase Phlebioid fungi include the model species of white rot encoding genes, and their deduced protein products decay of wood, Phanerochaete chrysosporium [12]together Mäkinen et al. BMC Genomics (2019) 20:430 Page 14 of 22

oxidoreductases inherent among phlebioid fungi facilitates the attack on lignin and crystalline cellulose which further- more allows access to the more easily degradable and utilizable plant biomass polysaccharides. Of the phlebioid species selected for this comparative study, P. chrysosporium has been the most extensively studied and was the first genome of Basidiomycota fun- gus to be sequenced in early 2000 [12]. Since then, hun- dreds of Basidiomycota including over 80 species of the order Polyporales have been genome sequenced in the 1000 Fungal Genomes project (JGI MycoCosm https:// genome.jgi.doe.gov/programs/fungi/index.jsf)[16, 49]. The sequenced and annotated phlebioid clade species of Polyporales comprise one or a few species of the genera Phanerochaete, Phlebia, Phlebiopsis, and Bjerkandera. Although the phlebioid fungi are deadwood and woody debris colonizing and decomposing species, they deviate in their ecological and physiological characteristics. P. chrysosporium has been mainly isolated from hardwoods (angiosperm wood) whereas Phanerochaete carnosa usu- ally inhabits softwood (coniferous wood) [29]. Phlebiop- sis gigantea is considered as a pioneer colonizer of softwood due to its special ability to grow into freshly ex- Fig. 7 Venn diagram illustrating the number of shared SSP posed sapwood, such as cut tree stumps, with a candidates of P. radiata. The SSP genes were computationally high content of resinous extractives, which apparently re- analysed by using EffectorP and LOCALIZER and according to the criteria of being small and cysteine-rich stricts substrate access for other wood-decay fungi [15]. Bjerkandera adusta in turn is a common Polyporales spe- cies in temperate European forests inhabiting dead angio- with the efficient lignin-attacking species Phlebia radiata sperm hardwood, such as beech, and coniferous wood. B. [18, 26] and xenobiotics modifying species Bjerkandera adusta is considered as a secondary or primary colonizer adusta [7, 47], all of which are well-known producers of a of attached branches, tree stumps and fallen wood [50]. large array of secreted enzymes against wood carbohydrates Phlebia species, on the other hand, are likewise sec- (CAZymes) including numerous auxiliary oxidoreductases ondary or primary colonizers able to cause white rot (oxidases, laccases, peroxidases and monooxygenases) [5, both on coniferous and hardwood, with some substrate 13, 33]. Especially, production of the oxidoreductases to specificities [51, 52]. The resupinate fruiting bodies of P. generate extracellular ROS (reactive oxygen species) [48] radiata are commonly found on dead angiosperm wood and to modify lignin by production of lignin-active class-II such as stumps and branches of birch and rowan, peroxidases [13, 14, 37] are characteristics of the phlebioid whereas P. centrifuga is a common colonizer of dead co- species. The strong ability for production of ROS and nifers in the boreal forests of North Europe [51]. P.

Table 7 Summary of the SSP candidates of P. radiata that fulfilled the three criteria used in the analyses Protein coding gene ID Length Number of cysteines EffectorPa LOCALIZERb minus.g2681 265 10 0.610 NLS minus.g6180 200 6 0.669 0.998, C plus.g1485 119 11 0.916 NLS plus.g2909 150 8 0.676 0.998, C plus.g2957 146 8 0.570 0.896, C plus.g7370 229 19 0.767 0.716, C plus.g8107 158 8 0.747 0.971, C; 0.824, M plus.g881 179 4 0.781 0.993, C; 0.995, M aThe number is the probability of being an effector given by EffectorP bThe number is the probability of having a Chloroplast (C) transit peptide-like or Mitochondria (M) transit peptide-like sequences given by LOCALIZER; for the cases of searching for nuclear localization signals (NLS), no probability was returned Mäkinen et al. BMC Genomics (2019) 20:430 Page 15 of 22

brevispora was described as a new species causing sub- for the white rot phlebioid fungi for efficient decom- stantial white rot decay in pine wood in North America position of wood lignocellulose. [52]. Thus, by the aid of comparative genomic analyses The CAZy family AA9 lytic polysaccharide monooxy- on the wood carbohydrate and lignin active CAZyme en- genases attack and cleave polymeric cellulose via oxida- coding genes, we expected to observe some differences in tive and oxygen atom incorporating mechanism [48, 58, the gene repertoire between the seven species of taxonom- 59] and may also act on hemicelluloses [60]. The AA9 ically near-related and eco-physiologically specified phle- lpmo genes are expanded in white rot Polyporales fungi bioid white-rot fungal species. [57], and accordingly, the studied phlebioid genomes As an outcome of the complete sequencing, assem- contained over 10 annotated lpmo AA9 genes. Especially bly, gene prediction and annotation of the P. radiata notable is the large number of AA9 family genes (up to 28 nuclear genome, a fully closed and gene dense refer- genes) in B. adusta. The novel auxiliary enzyme LPMO ence genome was obtained. Most of the sequenced family AA14 encoding genes were additionally identified phlebioid genomes are draft assemblies and there has in the seven phlebioid genomes, indicative of enzyme ac- been a need for a more accurate reference genome, tivities necessary for efficient white rot fungal lifestyle, al- which will now be filled by the availability of P. though the number of genes (2-3) was much lower than radiata genome. The predicted gene number of P. annotated for the AA9 family LPMOs. The family AA14 radiata is the second highest among the phlebioid ge- LPMOs were recently discovered to present oxidative re- nomes. Specific features of the predicted protein- action against the heteroxylan shield covering cellulose coding genes in P. radiata were the extremely short microfibrils [35], thus potentially facilitating access to cel- exons, even as short as two nucleotides, and many lulose backbone for the cellulolytic enzymes. exons comprising only one codon (three nucleotides). Taking into account the recent finding of activation of For example, the glycolytic core metabolic GAPDH LPMOs preferentially by hydrogen peroxide (rather than enzyme encoding gene begins with the first start dioxygen) [48, 61], and the requirement of hydrogen per- codon exon followed by the immediate first intron oxide as the primary oxidant for the CAZy family AA2 (gene plus.g5457). In addition, the 5´ UTRs of the an- lignin-attacking class-II peroxidases [5, 37], enzymatic notated genes of P. radiata typically contain putative production of H2O2 and other ROS for biological decom- upstream open reading frames (uORFs). In fungal position of lignocellulose is of utmost importance. Of the mRNAs, uORFs are fairly common and are believed various auxiliary oxidoreductase enzymes, the potential to have for example regulatory roles for gene expres- hydrogen peroxide producing GMC flavoprotein family sion and possibly an effect on mRNA stability [53]. AA3 is presented with 23-39 genes in the phlebioid fungi, Exploring the CAZyme gene content among the and all the studied genomes harbor one cdh gene encod- seven phlebioid fungi revealed gene sets typical for ing a cellobiose dehydrogenase (CDH, AA3_1). The spe- white rot Basidomycota fungi of the class Agaricomy- cific function of the hybrid flavo-hemoprotein CDH is still cetes [5, 7, 8, 13, 28, 30] encoding an array of se- unclear; in addition to possibly participating in lignin con- creted cellulases, hemicellulases, pectinases, auxiliary version, the enzyme may be involved in attack on cellulose oxidoreductases and lignin-modifying enzymes. In and hemicellulose in cooperation with LPMOs [62, 63]. addition to the conventional cellobiohydrolase en- A wide range of activities are needed for the degrad- coding genes of the families GH6 (1 gene in phle- ation of the complex hemicellulose polymers [64, 65]. bioid genomes) and GH7 (3-8 genes in each fungus), Endo-xylanases and mannanases cut the polymeric xylan all the studied genomes harbor GH131 genes belong- and mannan backbone, respectively, whereas mannosi- ing to a novel CAZy family, which includes unusual dases and xylosidases release monomeric sugars. For ex- hydrolases with bifunctional exo-β-1,3-/-1,6- and ample, acetyl xylan esterases, acetyl esterases, glucuronyl endo-β-1,4 activities towards a wide range of β- esterases, glucuronidases, galactosidases, fucosidases and glucans including also cellulosic derivatives [54]. arabinofuranosidases are needed for releasing the vari- Another unconventional cellulase encoding gene ous linkages and side chains moieties present in chem- present in the phlebioid genomes belongs to the ically and structurally diverse hemicelluloses [64]. Genes family GH9. The P. chrysosporium GH9 gene has encoding all activities necessary for hemicellulose de- been cloned and characterized [55] and GH9 enzyme composition were detected in the studied phlebioid ge- of the bacterium Clostridium cellulosi was recently nomes, although some differences in the set of genes described as a processive endoglucanase [56]. It is of were observed. For example, the CAZy family GH29 α- interest that the brown rot fungal genomes of Poly- fucosidase encoding genes were present only in the Phle- porales seemingly lack GH9 genes [57]. Thus, it may bia species, and interestingly, these enzymes could be be proposed that the higher degree of hydrolytic en- important in the early phases of wood colonization [9, zyme diversity against cellulose gives an advantage 43]. Furthermore, the family GH11 endoxylanases may Mäkinen et al. BMC Genomics (2019) 20:430 Page 16 of 22

imply specific functions due to their uneven distribution fungi, which may imply concerted regulation of gene ex- as corresponding genes among the phlebioid genomes. pression [66]. In the case of P. gigantea, the lack of GH95 α-fucosidase Phylogenetic conservation of specific gene clusters pin- encoding genes (1-3 genes in other phlebioid genomes) points their importance in the fitness for fungi in their en- as well as the abundance of CE8 pectin methylesterase vironments, and thereby, preservation in the course of encoding genes were noticed in the previous genome de- evolution [67]. Clustering of lignin-modifying peroxidase scription study [15]. and copper radical oxidase genes of P. chrysosporium has The phlebioid genomes possess key genes encoding ac- been recognized previously [12], whereas clustering of cel- tivities against lignin including genes of AA2 family class lulase encoding genes was less evident. Similarly, cluster- II peroxidases together with genes of AA3, AA4, AA5 ing of the P. carnosa MnP and LiP encoding genes has families encoding proteins for production of hydrogen been observed, whereas most of the CAZyme genes were peroxide. The genera Phlebia, Phanerochaete and Bjer- loosely associated [29]. In P. radiata, we noticed in this kandera present various types of MnP and LiP enzymes, study similar features of clustering of LiP encoding genes including one VP in B. adusta [14], whereas Phlebiopsis (two genes) and likewise noticed in P. chrysosporium,to- gigantea genome is restricted to short- and atypical gether with AA5 copper radical oxidase encoding genes MnPs [15]. It is evident that the class-II heme peroxi- (three genes). For gene clustering and co-expression, sev- dases are necessary for oxidation and modification of lig- eral explanations may be proposed. First, the neighboring nin [37], and the appearance of the first class-II genes may locate in the genome on the same euchromatin peroxidase gene in the ancestor of Agricomycetes 290 or heterochromatin segment, or the co-expressed genes million years ago initiated decomposition of wood lignin may contain a common transcription factor binding site and white rot lifestyle [13]. (on the promoter region) or even share the same pro- Notably, expansion in the number of AA2 class-II per- moter [66]. However, as clustering of genes is not neces- oxidase encoding genes (10-21 genes per genome) were sary for co-expression, the true purpose of gene clustering annotated with several enzymes produced and secreted on may be positioning the physiologically important and wood substrates in the most efficient white rot colonizers beneficial genes into chromosomal areas of low recombin- and lignin-degrading fungi (Phanerochaete chrysosporium, ation rate in the fungal genomes [38, 67]. P. carnosa, Phlebia radiata, Bjerkandera adusta)[12, 13, A holistic study of the transcriptome and proteome of P. 26, 29, 57]. The expansion of AA2 lignin-modifying perox- radiata was recently conducted, which confirmed that all idase and dye-decolorizing peroxidase (DyP) to 21 and 10 proteinsnecessaryforthewhite-rottypeofdecayarepro- genes, respectively, in B. adusta was previously docu- duced on solid spruce wood, and their corresponding genes mented [14]. The evolutionary analysis suggested that the were up-regulated [26].Thetimescaleanalysisofproteo- ancestor of Agaricomycetes possessed one to two genes for mics revealed that several lignin-modifying class-II peroxi- class-II peroxidases, and one to two genes coding for dye- dases together with glyoxal and alcohol oxidases were decolorizing peroxidases [13]. The number of DyP encod- abundantly produced at earlier growth stage on wood, indi- ing genes in the phlebioid fungi has expanded from one to cating an initial oxidative attack against lignin units. Dy- four genes in P. gigantea and to ten genes in B. adusta, namic changes in quantities of especially LiP and MnP whereas the efficiently lignin-attacking species P. chrysos- enzymes were detected during the six-week cultivation on porium has evidently lost this gene. Thus, it may be con- wood. P. radiata is also able to activate its wood decompos- cluded that on the contrary to the class-II LiP, MnP and ing enzymatic machinery under oxygen-limited cultivation VP peroxidases, DyP enzymes are not essential for oxi- atmosphere in order to bioconvert lignocellulosic waste dization and degradation of lignin. materials while simultaneously producing ethanol and Overall, the CAZyme gene content of the studied fun- other metabolites [68, 69]. gal genomes was quite constant, with over 150 anno- Of the multitude of protein-degradative activity (peptid- tated genes per genome, which is expected from closely ase) encoding genes recognized in P. radiata genome, se- related fungi with similar, white-rot decay and wood creted peptidases produced during the six-week cultivation inhabiting lifestyles. However, some fine-tuning of plant on spruce wood were previously identified in the proteome cell wall degradation activities may be proposed from [26], and the abundance of peptidases was shown to in- the subtle differences detected, especially in the number crease over time. Together with an increase in the abun- of CAZyme genes coding for enzyme activities against dance of peptidases in the proteome on wood, the hemicelluloses. In addition, the family GH44 endogluca- simultaneous increase of e.g. chitinases [26] indicated that nase encoding genes were found to be specific for the the fungal cell wall is undergoing reorganization and essen- three Phlebia species. Moreover, genomic clustering of tial nutrients, predictably mainly nitrogen compounds, especially AA9 lpmo genes and GH10 xylanase encoding were recycled upon growth on wood. The most abundant genes was noticed to be common among the phlebioid peptidases detected belong to the A01A and T01A Mäkinen et al. BMC Genomics (2019) 20:430 Page 17 of 22

subfamilies according to the MEROPS classification. In this mask their recognition by mammalian immune system, study, annotation of the peptidase encoding genes revealed thereby rendering silencing of immune response [79]. In that P. radiata genome encodes 31 and 11 putative mem- the rice blast Ascomycota fungus Magnaporthe grisea, bers of the peptidase A01A and T01A families, respect- hydrophobins have been considered as pathogenicity fac- ively, which explains their dominance in the expressed tors involved in conidial development and viability, ap- proteome on wood. The most abundant peptidases identi- pressorium formation, infectious growth in host cells fied from the proteome included a M28 metalloprotease, and possibly in cell wall assembly [80]. Our data could an A01 aspartyl peptidase and an S53 tripeptidyl peptidase provide an example of more roles of hydrophobins in [26]. the Basidiomycota and in general, in saprotrophic fungi. The diversity of genes potentially involved in biosynthesis Analysis of putative ABC transporters indicated that of secondary metabolites (SM) observed in P. radiata is the set of P. radiata ABC genes identified in the genome quiteuncommonforawhiterotPolyporales and class is characteristic for the species of the Basidiomycota Agaricomycetes fungal species [8, 15], which is an indication subphylum Agaricomycotina [41, 42]. A notable feature of the ability of P. radiata to produce a diverse set of sec- of the set of ABC genes in P. radiata is a high number ondary metabolites. Altogether, genes for polyketide syn- of genes coding ABCC4 transporters, particularly thase, adenylate-forming reductase, terpene cyclase, and present in eight copies in the most expanded ABC-C putative AMP-dependent acyl-CoA synthetase were identi- subfamily, but all genes and their protein products of yet fied and annotated in the genome of P. radiata.Protein unknown function. Previous phylogenetic analysis indi- family of adenylate-forming reductases is present in the Ba- cates that numerous lineage-specific duplication events sidiomycota species, but the functions of these enzymes in of this gene occurred in the evolution of Agaricomyco- fungi are not entirely understood [70]. Clustering of the tina [41], but their functional significance remains ob- secondary metabolism genes is commonly observed in scure. The subfamily of ABC-C transporters of unknown fungi, which may contribute to the regulation of genes be- function, are depicted in great variation in copy number longing to a single SM pathway [71]. The occurrence of among Agaricomycotina. The gene is present in a single secondary metabolism genes close to CAZyme genes, copy in about one third of the analyzed species, but in which was observed in P. radiata for a few occasions, has some species, in particular in the members of Polypor- likewise been noted in the Ascomycota species Trichoderma ales, more than 10 genes are recognized [41], which fur- reesei although the possible biological reason for this coin- thermore implies the necessity of a versatile set of cidence is not yet known [38, 72]. transporter proteins for saprotrophic lifestyle wood- Effectors secreted by fungal plant pathogens during decay fungi to enable efficient nutrient uptake and meta- colonization of the host are well known examples of small bolic activity in their harsh habitats. secreted proteins (SSPs) [73, 74]. However, the role of SSPs In conclusion, it may be inferred that the genome of in the lifestyle of saprotrophic fungi such as in the seven Phlebia radiata undoubtedly places the fungus and spe- wood decay phlebioid species of this study, is still enig- cies among the phlebioid clade of Polyporales, class matic. A recent study suggests that some of the SSPs identi- Agaricomycetes, within the fungal phylum Basidiomy- fied in Basidiomycota genomes could participate in the cota. Especially the comparative genomic analyses of the lignin-attacking system [75]. Other possible roles for SSPs sets of CAZyme and auxiliary oxidoreductase enzyme are development of multicellular structures, and coping encoding genes among the seven phlebioid fungal ge- with toxic compounds present in the environment [76]or nomes emphasize the significance of the wide array of perhaps, involvement in the interspecific fungal interac- genes for activities against all polymeric components of tions. For example, one of the P. radiata genes predicted to plant cell wall (cellulose, hemicelluloses, pectin, lignin) encode a SSP (plus.g7370) is homologous with the priA for a saprotrophic, white rot type of wood decay strat- gene Lentinula edodes for which a role during the early egy, and successful lifestyle in the forest environments. stage of fruiting body formation has been proposed [77]. Importantly, many CAZyme genes were clustered in One intriguing notion previously was the high expres- the genome, and interestingly, some CAZyme genes sion level of hydrophobin-coding genes in P. radiata show near location with secondary metabolism biosyn- during colonization of spruce wood [26]. Hydrophobins thesis genes, which suggests potential co-expression and are a large family of secreted proteins unique to fila- functional roles for the produced metabolites in inhabit- mentous fungi [78]. These proteins function by forming ation and decomposition of wood by P. radiata. New into polymeric, water-repellent monolayers at the sur- findings on the abundance and versatility of the ABC face of aerial structure such as spores, hyphae and fruit- transporter and small secreted protein encoding genes ing bodies [78]. The surface (rodlet) layer coating the will aid in identifying them more extensively in the fun- conidial spores of the pathogenic Ascomycota filament- gal genomes and transcriptomes. Our comparative ap- ous fungal species Aspergillus fumigatus was found to proach and bioinformatic results are adding up the Mäkinen et al. BMC Genomics (2019) 20:430 Page 18 of 22

fundamental knowledge on fungal genomics and biology, by blastp (version 2.2.30) [85] searches against the NCBI which may be employed in research on fungal metabolic non-redundant protein sequences database, by PANNZER processes and bioconversions, as well as a source for [86] and by Blast2GO [87] software. Gene models of inter- new genes and proteins for biotechnology applications. est were manually curated.

Methods GO enrichment analysis Fungal strain and isolation of DNA For functional prediction, GO enrichment analysis based Phlebia radiata Fr. isolate 79 (FBCC0043) was originally on Fisher´s exact test was conducted by using Blast2GO isolated from a decaying grey alder (Alnus incana) trunk [87] with data from our previous transcriptome study in South Finland, and is deposited in the Microbial Domain [26]. Filter mode was p-value of 0.05 for statistical sig- Biological Resource Centre HAMBI (HAMBI-FBCC sub- nificance. Genes up-regulated (p-value < 0.05, log2 FC ≥ collection, kotka.luomus.fi/culture/fbcc) of the University of 1) upon growth on spruce wood as compared with malt Helsinki. Identification of the isolate has been verified by extract medium as substrate at the two-week and/or ITS-PCR [10, 81]. The isolate was cultivated and main- four-week time points of cultivation were selected for tained on malt extract (2 % w/v) agar at 25 °C and in the the analysis. dark throughout the study. For isolation of total DNA, P. radiata wascultivatedinliquidmaltextract(2%w/v) mediumfor7-14daysat25°Cinthedarkafterwhichthe Annotation and comparison of CAZyme and peptidase mycelium was harvested, frozen to -20 °C and lyophilized. encoding genes Isolation of genomic DNA was initiated with CTAB- P. radiata CAZyme genes were functionally annotated supplemented incubation of fine ground mycelium followed previously [26]. Majority of the CAZyme encoding genes by phenol-chloroform purification, RNAse treatment, and involved in decomposition of plant cell wall lignocellulose final precipitation in pure cold ethanol [82]. were manually curated. Information on annotation of pu- tative P. radiata peptidases according to the MEROPS Genome sequencing, assembly and gene prediction peptidase database [40](https://www.ebi.ac.uk/merops/) Prior to sequencing, purity and molecular size distribution was collected from the Joint Genome Institute (JGI) of total DNA (2.0 mg) was analysed by Qubit (Thermo MycoCosm fungal genomics repository [16](https://gen- Scientific) and Bioanalyzer (Agilent Technologies). PacBio ome.jgi.doe.gov/programs/fungi/index.jsf). For compara- single molecule, real-time (SMRT) sequencing technology tive genomics of the CAZyme and peptidase genes, (Pacific Biosciences of California) was the basic method genetic information of the phlebioid fungi Phlebia brevis- used to obtain long nucleotide reads for de novo assembly pora (HHB-7030 v1.0), Phlebia centrifuga (FBCC195), of the genome. In total, 9 SMRT cells were run which Phlebiopsis gigantea (11061_1 CR5-6 v1.0), Bjerkandera yielded 70 2780 reads with N50 read length of 9 384 bp adusta (HHB-12826-SP v1.0), Phanerochaete chrysospor- and a total of 4.29 Gbases of nucleotide sequence data. ium (RP-78 v2.2) and Phanerochaete carnosa (HHB- HGAP3 from SMRT Analysis v2.3 suite was used with 6 10118-Sp v1.0) was downloaded from the JGI MycoCosm. 000 bp seed read length to assemble the genome. The as- sembly was polished with three subsequent runs of SMRT Genes for secondary metabolism and ABC transporters Analysis re-sequencing pipeline with Quiver consensus al- Identification of the secondary metabolism-related genes gorithm. Low coverage (<5X) and redundant, such as in P. radiata genome was performed using antiSMASH ribosomal only, short contigs were removed. Strand- prediction platform [88] with the default settings for specific paired-end Illumina RNA-Seq data described in fungal genomes followed by manual curation and anno- [26] was mapped against the final assembly of genome se- tation of the identified genes. P. radiata ABC protein- quence using the splice-aware RNA-Seq aligner STAR coding genes were identified by multiple tblastn and [83]. Mapped RNASeq data was filtered using the follow- blastp searches against the genome assembly and a set ing parameters: 1) both R1 and R2 were properly mapped, of predicted proteins, respectively. Sequences of previ- 2) insert size between the reads was more than 500 bp, ously identified P. brevispora ABC proteins [41] were and 3) the mapping quality was 255. The cleaned RNA- used as queries. All hits producing E-values below 10-6 Seq bam file was finally divided into plus and minus were further examined. Gene models were manually cu- strand read pairs based on the strand-specific nature of rated and, when necessary, the positions of N and C ter- the libraries (R2 data corresponding the gene transcription mini were adjusted. For gene number comparisons, the orientation) before genome annotation. Coding sequences number of ABC transporter genes of the studied phle- of protein-coding gene models were predicted by the bioid fungi was obtained from the previous study [41]. BRAKER1 software [84] using the plus and minus strand The putative ABC transporter genes of P. centrifuga RNA-Seq evidence separately, and functionally annotated were then identified by blastp searches against the Mäkinen et al. BMC Genomics (2019) 20:430 Page 19 of 22

genome sequence deposited in the MycoCosm with the proteome data from spruce wood cultivations [26]. 2w, two week time P. radiata ABC proteins as queries. point of the spruce wood cultivation; 4w, four week time point of the spruce wood cultivation; up, up-regulated; down, down-regulated. (XLSX 44 kb) Small secreted proteins Additional file 3: P. radiata gene models encoding plant cell wall Secreted proteins were identified using SignalP v.4.1 attacking CAZymes and located in clusters in the genome assembly. Fold (sensitive mode) [89], TargetP v.1.1 [90] and TMHMM change of expression of the genes in spruce wood cultivations as v.2.0 [91] to predict the presence of signal peptide, tar- compared with malt extract cultivations at the two week (2w) and four week (4w) time points of cultivation was analyzed previously [26]. (XLSX geted cellular localization and transmembrane domain 22 kb) (TM), respectively. Proteins having more than two TMs Additional file 4: Putative secondary metabolism genes identified from and or having a single TM not overlapping with the sig- the P. radiata genome assembly. Fold change of expression of the genes nal peptide (outside the first 60 amino acids) were ex- in spruce wood cultivations as compared with malt extract cultivations at the two week (2w) and four week (4w) time points of cultivation was cluded. Subsequently, the predicted secretome was analyzed previously [26]. (XLSX 15 kb) submitted to EffectorP v.2.0 [92] and LOCALIZER v.1.0 Additional file 5: Putative small secreted proteins of P. radiata chosen [93]. EffectorP could help prioritize the selection of according to EffectorP analysis, LOCALIZER analysis and by identifying highly confident effector candidates. LOCALIZER aids small cysteine rich protein. Fold change of expression of the genes in spruce wood cultivations as compared with malt extract cultivations at the prediction of translocation of fungal effectors into the two week (2w) and four week (4w) time points of cultivation was plant chloroplasts, mitochondria and nuclei based on analyzed previously [26]. (XLSX 40 kb) the presence of transit peptides and nuclear localization Additional file 6: Distribution of length of Phlebia radiata whole signals. In addition, secreted proteins with length < 300 proteome, secretome and SSPs. Numbers on the bar denote the average length of each set of protein sequences. (PDF 19 kb) amino acids and cysteine residues ≥ 4 were also selected. Additional file 7: ABC transporter genes identified from the P. radiata All three sets of putative proteins were merged and con- genome assembly. (XLSX 117 kb) stituted the final SSP list after removing the putative se- Additional file 8: Comparisons of the numbers of ABC transporter creted CAZymes. genes between different species of Agaricomycotina. (XLSX 23 kb)

Recognition of genomic clustering and co-expression of Abbreviations CAZyme genes AA: Auxiliary activity redox enzyme; CAZyme: Carbohydrate-active enzyme; CE: Carbohydrate esterase; GH: Glycoside hydrolase; LiP: Lignin peroxidase; Co-localization of P. radiata CAZyme genes in the gen- LPMO: Lytic polysaccharide monooxygenase; MnP: Manganese peroxidase; ome assembly were searched according to unitig number NRPS: Non-ribosomal peptide synthase; nt: Nucleotide; PKS: Polyketide and nucleotide start and end positions of the predicted synthase; PL: Polysaccharide lyase; SM: Secondary metabolism gene models. Minimum of three CAZyme genes in a Acknowledgements window of maximum of 11 genes was considered as a Outi-Maaria Sietiö is thanked for the aid in purification of P. radiata DNA for cluster. Size of the window was calculated according to genome sequencing. Igor Grigoriev at DOE JGI, USA, is thanked for collaboration and accepting the annotated genome assembly as submission [38]. Common expression patterns i.e. up-regulation at 2 to the MycoCosm repository. and 4 weeks of growth on spruce wood substrate as compared with malt extract liquid medium were identi- Authors’ contributions TL, JK, LP, PA and FOA designed the study. JK performed fungal cultivations fied [26]. Co-localization of the GH10 and AA9 family and DNA and RNA extractions. LP managed sequencing and data collection. genes of P. brevispora, P. centrifuga, P. gigantea, B. adu- OPS and PL assembled P. radiata genome and performed gene model sta, P. chrysosporium and P. carnosa was investigated prediction. MM and JK annotated and manually curated CAZyme gene models, and performed comparative protein analyses. MM analyzed GO with the graphical display of the JGI Genome Browser enrichment and genomic clustering. AK annotated the secondary (https://genome.jgi.doe.gov/help/browser_main.jsf)bymetabolism and ABC transporter genes. ZZ annotated the small secreted visualizing the clusters of the CAZyme gene models po- protein genes. TL, PA and FOA supported materials. MM and TL wrote and sitioned along each genome assembly scaffold. Max- edited the manuscript. All authors read and approved the final manuscript. imum number of three non-CAZyme genes was allowed Funding between the genes forming the cluster. This study was financed by the Academy of Finland project Grants No. 138331 and 285676 to TL, and 276862 to FOA.

Additional files Availability of data and materials The annotated P. radiata genome assembly is available at the Joint Genome Additional file 1: GO enrichment analysis of P. radiata genes up- Institute (JGI) MycoCosm repository (https://genome.jgi.doe.gov/Phlrad1/ regulated in the presence of spruce wood. Statistically significant up- Phlrad1.home.html). All data generated and analysed in the study are regulation (p-value < 0.05, log2 FC ≥ 1) was analyzed previously [26]as included in this published article and its supplementary information files. compared with malt extract cultivation at the two week (2w up- regulated) and/or four week (4w up-regulated) time points of cultivation. Ethics approval and consent to participate (XLSX 11 kb) Not applicable. Additional file 2: P. radiata gene models encoding CAZymes, their locations in the genome assembly together with transcriptome data and Consent for publication Not applicable. Mäkinen et al. BMC Genomics (2019) 20:430 Page 20 of 22

Competing interests wood. PLoS Genet. 2014;10:e1004759. https://doi.org/10.1371/journal. The authors declare that they have no competing interests. pgen.1004759. 16. Grigoriev IV, Nikitin R, Haridas S, Kuo A, Ohm R, Otillar R, et al. Author details MycoCosm portal: gearing up for 1000 fungal genomes. Nucleic Acids 1Department of Microbiology, Faculty of Agriculture and Forestry, Viikki Res. 2014;42(Database issue):D699–704. https://doi.org/10.1093/nar/ Campus, University of Helsinki, FI-00014 Helsinki, Finland. 2DNA Sequencing gkt1183. and Genomics Laboratory, Institute of Biotechnology, Viikki Campus, FI-00014 17. Justo A, Miettinen O, Floudas D, Ortiz-Santana B, Sjökvist E, Lindner D, Helsinki, Finland. 3Department of Forest Sciences, Faculty of Agriculture and et al. A revised family-level classification of the Polyporales Forestry, University of Helsinki, Viikki Campus, FI-00014 Helsinki, Finland. (Basidiomycota). Fungal Biol. 2017;121:798–824. https://doi.org/10.1016/J. 4Present address: VTT Technical Research Centre of Finland Ltd., Espoo, FUNBIO.2017.05.010. Finland. 5Present address: Department of Chemistry and Biotechnology, 18. Lundell T, Leonowicz A, Rogalski J, Hatakka A. Formation and action of Division of Gene Technology, Tallinn University of Technology, Tallinn, lignin-modifying enzymes in cultures of Phlebia radiata supplemented with Estonia. veratric acid. Appl Envir Microbiol. 1990;56:2623–9 http://aem.asm.org/ content/56/9/2623.short. Accessed 26 Jun 2015. Received: 7 March 2019 Accepted: 21 May 2019 19. Mäkelä MR, Lundell T, Hatakka A, Hildén K. Effect of copper, nutrient nitrogen, and wood-supplement on the production of lignin-modifying enzymes by the white-rot fungus Phlebia radiata. Fungal Biol. 2013;117:62– 70. https://doi.org/10.1016/j.funbio.2012.11.006. References 20. Hakala TK, Maijala P, Konn J, Hatakka A. Evaluation of novel wood-rotting 1. Boerjan W, Ralph J, Baucher M. Lignin biosynthesis. Annu Rev Plant Biol. polypores and for the decay and biopulping of Norway 2003;54:519–46. https://doi.org/10.1146/annurev.arplant.54.031902.134938. spruce (Picea abies) wood. Enzyme Microb Technol. 2004;34:255–63. https:// 2. Gillet S, Aguedo M, Petitjean L, Morais ARC, da Costa Lopes AM, Łukasik RM, doi.org/10.1016/j.enzmictec.2003.10.014. et al. Lignin transformations for high value applications: towards targeted 21. Hofrichter M, Lundell T, Hatakka A. Conversion of milled pine wood by modifications using green chemistry. Green Chem. 2017;19:4200–33. https:// manganese peroxidase from Phlebia radiata. Appl Environ Microbiol. 2001; doi.org/10.1039/C7GC01479A. 67:4588–93. https://doi.org/10.1128/AEM.67.10.4588-4593.2001. 3. Hibbett DS, Binder M, Bischoff JF, Blackwell M, Cannon PF, Eriksson OE, et al. 22. Hildén KS, Mäkelä MR, Hakala TK, Hatakka A, Lundell T. Expression on A higher-level phylogenetic classification of the Fungi. Mycol Res. 2007;111: wood, molecular cloning and characterization of three lignin 509–47. https://doi.org/10.1016/J.MYCRES.2007.03.004. peroxidase (LiP) encoding genes of the white rot fungus Phlebia 4. McLaughlin DJ, Hibbett DS, Lutzoni F, Spatafora JW, Vilgalys R. The search radiata. Curr Genet. 2006;49:97–105. https://doi.org/10.1007/s00294- for the fungal tree of life. Trends Microbiol. 2009;17:488–97. https://doi.org/ 005-0045-y. 10.1016/J.TIM.2009.08.001. 5. Lundell TK, Mäkelä MR, de Vries RP, Hildén KS. Genomics, lifestyles and 23. Hildén K, Martínez AT, Hatakka A, Lundell T. The two manganese future prospects of wood-decay and litter-decomposing Basidiomycota. Adv peroxidases Pr-MnP2 and Pr-MnP3 of Phlebia radiata, a lignin-degrading Bot Res. 2014;70:329–70. Fungi. Elsevier. https://doi.org/10.1016/B978-0-12- basidiomycete, are phylogenetically and structurally divergent. Fungal – 397940-7.00011-2. Genet Biol. 2005;42:403 19. https://doi.org/10.1016/j.fgb.2005.01.008. 6. Baldrian P. Microbial activity and the dynamics of ecosystem processes in 24. Mäkelä MR, Hildén KS, Hakala TK, Hatakka A, Lundell TK. Expression and forest soils. Curr Opin Microbiol. 2017;37:128–34. https://doi.org/10.1016/J. molecular properties of a new laccase of the white rot fungus Phlebia – MIB.2017.06.008. radiata grown on wood. Curr Genet. 2006;50:323 33. https://doi.org/10. 7. Binder M, Justo A, Riley R, Salamov A, Lopez-Giraldez F, Sjökvist E, et al. 1007/s00294-006-0090-1. Phylogenetic and phylogenomic overview of the Polyporales. Mycologia. 25. Saloheimo M, Niku-Paavola M-L, Knowles JKC. Isolation and structural 2013;105:1350–73. https://doi.org/10.3852/13-003. analysis of the laccase gene from the lignin-degrading fungus Phlebia – 8. Riley R, Salamov AA, Brown DW, Nagy LG, Floudas D, Held BW, et al. radiata. J Gen Microbiol. 1991;137:1537 44. https://doi.org/10.1099/ Extensive sampling of basidiomycete genomes demonstrates inadequacy of 00221287-137-7-1537. the white-rot/brown-rot paradigm for wood decay fungi. Proc Natl Acad Sci 26. Kuuskeri J, Häkkinen M, Laine P, Smolander OP, Tamene F, Miettinen S, et al. U S A. 2014;111:9923–8. https://doi.org/10.1073/pnas.1400592111. Time-scale dynamics of proteome and transcriptome of the white-rot 9. Kuuskeri J. Genomics and systematics of the white-rot fungus Phlebia fungus Phlebia radiata: growth on spruce wood and decay effect on radiata: Special emphasis on wood-promoted transcriptome and proteome. lignocellulose. Biotechnol Biofuels. 2016;9:192. https://doi.org/10.1186/ University of Helsinki; 2016. s13068-016-0608-9. 10. Kuuskeri J, Mäkelä MR, Isotalo J, Oksanen I, Lundell T. Lignocellulose- 27. Mäkelä MR, Peng M, Granchi Z, Chin-A-Woeng T, Hegi R, van Pelt SI, et al. converting enzyme activity profiles correlate with molecular systematics and Draft genome sequence of the Basidiomycete white-rot fungus Phlebia phylogeny grouping in the incoherent genus Phlebia (Polyporales, centrifuga. Genome Announc. 2018;6. https://doi.org/10.1128/genomeA. Basidiomycota). BMC Microbiol. 2015;15:217. https://doi.org/10.1186/s12866- 01414-17. 015-0538-x. 28. Ohm RA, Riley R, Salamov A, Min B, Choi I-G, Grigoriev IV. Genomics of 11. Floudas D, Hibbett DS. Revisiting the of Phanerochaete wood-degrading fungi. Fungal Genet Biol. 2014;72:82–90. https://doi.org/10. (Polyporales, Basidiomycota) using a four gene dataset and extensive ITS 1016/j.fgb.2014.05.001. sampling. Fungal Biol. 2015;119:679–719. https://doi.org/10.1016/J.FUNBIO. 29. Suzuki H, MacDonald J, Syed K, Salamov A, Hori C, Aerts A, et al. 2015.04.003. Comparative genomics of the white-rot fungi, Phanerochaete carnosa and P. 12. Martinez D, Larrondo LF, Putnam N, Gelpke MDS, Huang K, Chapman J, et chrysosporium, to elucidate the genetic basis of the distinct wood types al. Genome sequence of the lignocellulose degrading fungus Phanerochaete they colonize. BMC Genomics. 2012;13:444. https://doi.org/10.1186/1471- chrysosporium strain RP78. Nat Biotechnol. 2004;22:695–700. https://doi.org/ 2164-13-444. 10.1038/nbt967. 30. Nagy LG, Riley R, Bergmann PJ, Krizsán K, Martin FM, Grigoriev IV, et al. 13. Floudas D, Binder M, Riley R, Barry K, Blanchette R, Henrissat B, et al. The Genetic bases of fungal white rot wood decay predicted by phylogenomic Paleozoic origin of enzymatic lignin decomposition reconstructed from 31 analysis of correlated gene-phenotype evolution. Mol Biol Evol. 2017;34:35– fungal genomes. Science. 2012;336:1715–9. https://doi.org/10.1126/science. 44. https://doi.org/10.1093/molbev/msw238. 1221748. 31. Kohler A, Kuo A, Nagy LG, Morin E, Barry KW, Buscot F, et al. Convergent 14. Ruiz-Dueñas FJ, Lundell T, Floudas D, Nagy LG, Barrasa JM, Hibbett DS, losses of decay mechanisms and rapid turnover of symbiosis genes in et al. Lignin-degrading peroxidases in Polyporales: an evolutionary mycorrhizal mutualists. Nat Genet. 2015;47:410–5. https://doi.org/10.1038/ survey based on 10 sequenced genomes. Mycologia. 2013;105:1428–44. ng.3223. https://doi.org/10.3852/13-059. 32. Aspeborg H, Coutinho PM, Wang Y, Brumer H, Henrissat B. Evolution, 15. Hori C, Ishida T, Igarashi K, Samejima M, Suzuki H, Master E, et al. substrate specificity and subfamily classification of glycoside hydrolase Analysis of the Phlebiopsis gigantea genome, transcriptome and family 5 (GH5). BMC Evol Biol. 2012;12:186. https://doi.org/10.1186/1471- secretome provides insight into its pioneer colonization strategies of 2148-12-186. Mäkinen et al. BMC Genomics (2019) 20:430 Page 21 of 22

33. Lombard V, Golaconda Ramulu H, Drula E, Coutinho PM, Henrissat B. The 54. Lafond M, Navarro D, Haon M, Couturier M, Berrin J-G. carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res. Characterization of a broad-specificity β-glucanase acting on β-(1,3)-, 2014;42:D490–5. https://doi.org/10.1093/nar/gkt1178. β-(1,4)-, and β-(1,6)-glucans that defines a new glycoside hydrolase 34. Horn S, Vaaje-Kolstad G, Westereng B, Eijsink VG. Novel enzymes for the family. Appl Environ Microbiol. 2012;78:8540–6. https://doi.org/10. degradation of cellulose. Biotechnol Biofuels. 2012;5:45. https://doi.org/10. 1128/AEM.02572-12. 1186/1754-6834-5-45. 55. Vanden Wymelenberg A, Denman S, Dietrich D, Bassett J, Yu X, Atalla R, et 35. Couturier M, Ladevèze S, Sulzenbacher G, Ciano L, Fanuel M, Moreau C, et al. Transcript analysis of genes encoding a family 61 endoglucanase and a al. Lytic xylan oxidases from wood-decay fungi unlock biomass degradation. putative membrane-anchored family 9 glycosyl hydrolase from Nat Chem Biol. 2018;14:306–10. https://doi.org/10.1038/nchembio.2558. Phanerochaete chrysosporium. Appl Environ Microbiol. 2002;68:5765–8. 36. Couturier M, Navarro D, Chevret D, Henrissat B, Piumi F, Ruiz-Dueñas FJ, et https://doi.org/10.1128/AEM.68.11.5765-5768.2002. al. Enhanced degradation of softwood versus hardwood by the white-rot 56. Zhang K-D, Li W, Wang Y-F, Zheng Y-L, Tan F-C, Ma X-Q, et al. Processive fungus Pycnoporus coccineus. Biotechnol Biofuels. 2015;8:216. https://doi.org/ degradation of crystalline cellulose by a multimodular endoglucanase via a 10.1186/s13068-015-0407-8. wirewalking mode. Biomacromolecules. 2018;19:1686–96. https://doi.org/10. 37. Hofrichter M, Ullrich R, Pecyna MJ, Liers C, Lundell T. New and classic 1021/acs.biomac.8b00340. families of secreted fungal heme peroxidases. Appl Microbiol Biotechnol. 57. Hori C, Gaskell J, Igarashi K, Samejima M, Hibbett D, Henrissat B, et al. 2010;87:871–97. https://doi.org/10.1007/s00253-010-2633-0. Genomewide analysis of polysaccharides degrading enzymes in 11 white- 38. Druzhinina IS, Kopchinskiy AG, Kubicek EM, Kubicek CP. A complete and brown-rot Polyporales provides insight into mechanisms of wood annotation of the chromosomes of the cellulase producer Trichoderma decay. Mycologia. 2013;105:1412–27. https://doi.org/10.3852/13-072. reesei provides insights in gene clusters, their expression and reveals genes 58. Hemsworth GR, Johnston EM, Davies GJ, Walton PH. Lytic polysaccharide required for fitness. Biotechnol Biofuels. 2016;9:75. https://doi.org/10.1186/ monooxygenases in biomass conversion. Trends Biotechnol. 2015;33:747–61. s13068-016-0488-z. https://doi.org/10.1016/j.tibtech.2015.09.006. 39. Huy ND, Le Nguyen C, Seo J-W, Kim D-H, Park S-M. Putative endoglucanase 59. Beeson WT, Vu VV, Span EA, Phillips CM, Marletta MA. Cellulose degradation PcGH5 from Phanerochaete chrysosporium is a β-xylosidase that cleaves by polysaccharide monooxygenases. Annu Rev Biochem. 2015;84:923–46. xylans in synergistic action with endo-xylanase. J Biosci Bioeng. 2015;119: https://doi.org/10.1146/annurev-biochem-060614-034439. 416–20 http://www.sciencedirect.com/science/article/pii/ 60. Agger JW, Isaksen T, Varnai A, Vidal-Melgosa S, Willats WGT, Ludwig R, et al. S1389172314003478?via%3Dihub. Discovery of LPMO activity on hemicelluloses shows the importance of 40. Rawlings ND, Barrett AJ, Finn R. Twenty years of the MEROPS database of oxidative processes in plant cell wall degradation. Proc Natl Acad Sci. 2014; proteolytic enzymes, their substrates and inhibitors. Nucleic Acids Res. 2016; 111:6287–92. https://doi.org/10.1073/pnas.1323629111. 44:D343–50. https://doi.org/10.1093/nar/gkv1118. 61. Müller G, Chylenski P, Bissaro B, Eijsink VGH, Horn SJ. The impact of 41. Kovalchuk A, Lee Y-H, Asiegbu FO. Diversity and evolution of ABC hydrogen peroxide supply on LPMO activity and overall proteins in basidiomycetes. Mycologia. 2013;105:1456–70. https://doi.org/ saccharification efficiency of a commercial cellulase cocktail. 10.3852/13-001. Biotechnol Biofuels. 2018;11:209. https://doi.org/10.1186/s13068-018- 42. Kovalchuk A, Kohler A, Martin F, Asiegbu FO. Diversity and evolution of ABC 1199-4. proteins in mycorrhiza-forming fungi. BMC Evol Biol. 2015;15:249. https://doi. 62. Henriksson G, Johansson G, Pettersson G. A critical review of cellobiose org/10.1186/s12862-015-0526-7. dehydrogenases. J Biotechnol. 2000;78:93–113. https://doi.org/10.1016/ 43. Olson Å, Aerts A, Asiegbu F, Belbahri L, Bouzid O, Broberg A, et al. Insight S0168-1656(00)00206-6. into trade-off between wood decay and parasitism from the genome of a 63. Courtade G, Wimmer R, Røhr ÅK, Preims M, Felice AKG, Dimarogona M, fungal forest pathogen. New Phytol. 2012;194:1001–13. https://doi.org/10. et al. Interactions of a fungal lytic polysaccharide monooxygenase with 1111/j.1469-8137.2012.04128.x. β-glucan substrates and cellobiose dehydrogenase. Proc Natl Acad Sci. 44. Kalb D, Lackner G, Hoffmeister D. Functional and phylogenetic divergence 2016;113:5922–7. https://doi.org/10.1073/pnas.1602566113. of fungal adenylate-forming reductases. Appl Environ Microbiol. 2014;80: 64. van den Brink J, de Vries RP. Fungal enzyme sets for plant polysaccharide 6175–83. https://doi.org/10.1128/AEM.01767-14. degradation. Appl Microbiol Biotechnol. 2011;91:1477–92. https://doi.org/10. 45. Christianson DW. Structural and chemical biology of terpenoid cyclases. Chem 1007/s00253-011-3473-2. Rev. 2017;117:11570–648. https://doi.org/10.1021/acs.chemrev.7b00287. 65. Rytioja J, Hildén K, Yuzon J, Hatakka A, de Vries RP, Mäkelä MR. Plant- 46. Mgbeahuruike AC, Kovalchuk A, Chen H, Ubhayasekera W, Asiegbu FO. polysaccharide-degrading enzymes from Basidiomycetes. Microbiol Mol Biol Evolutionary analysis of hydrophobin gene family in two wood-degrading Rev. 2014;78:614–49. https://doi.org/10.1128/MMBR.00035-14. basidiomycetes, Phlebia brevispora and Heterobasidion annosum s.l. BMC 66. Michalak P. Coexpression, coregulation, and cofunctionality of neighboring Evol Biol. 2013;13:240. https://doi.org/10.1186/1471-2148-13-240. genes in eukaryotic genomes. Genomics. 2008;91:243–8. https://doi.org/10. 47. Korniłłowicz-Kowalska T, Wrzosek M, Ginalska G, Iglik H, Bancerz R. 1016/J.YGENO.2007.11.002. Identification and application of a new fungal strain Bjerkandera adusta R59 67. Pál C, Hurst LD. Evidence for co-evolution of gene order and in decolorization of daunomycin wastes. Enzyme Microb Technol. 2006;38: recombination rate. Nat Genet. 2003;33:392–5. https://doi.org/10.1038/ 583–90. https://doi.org/10.1016/J.ENZMICTEC.2005.10.009. ng1111. 48. Bissaro B, Várnai A, Røhr ÅK, Eijsink VGH. Oxidoreductases and reactive 68. Mattila H, Kuuskeri J, Lundell T. Single-step, single-organism bioethanol oxygen species in conversion of lignocellulosic biomass. Microbiol Mol Biol production and bioconversion of lignocellulose waste materials by Rev. 2018;82:e00029–18. https://doi.org/10.1128/MMBR.00029-18. phlebioid fungal species. Bioresour Technol. 2017;225:254–61. https://doi. 49. Grigoriev IV, Nordberg H, Shabalov I, Aerts A, Cantor M, Goodstein D, et al. org/10.1016/j.biortech.2016.11.082. The genome portal of the Department of Energy Joint Genome Institute. 69. Mäkinen MA, Risulainen N, Mattila H, Lundell TK. Transcription of Nucleic Acids Res. 2012;40:D26–32. https://doi.org/10.1093/nar/gkr947. lignocellulose-decomposition associated genes, enzyme activities and 50. Hiscox J, Savoury M, Toledo S, Kingscott-Edmunds J, Bettridge A, Al Waili N, production of ethanol upon bioconversion of waste substrate by Phlebia et al. Threesomes destabilise certain relationships: multispecies interactions radiata. Appl Microbiol Biotechnol. 2018;102:5657–72. https://doi.org/10. between wood decay fungi in natural resources. FEMS Microbiol Ecol. 2017; 1007/s00253-018-9045-y. 93. https://doi.org/10.1093/femsec/fix014. 70. Brandenburger E, Braga D, Kombrink A, Lackner G, Gressler J, Künzler M, et 51. Kotiranta H, Saarenoksa R, Kytövuori I. Aphyllophoroid fungi of Finland. A al. Multi-genome analysis identifies functional and phylogenetic diversity of check-list with ecology, distribution, and threat categories. Norrlinia. 2009; basidiomycete adenylate-forming reductases. Fungal Genet Biol. 2018;112: 19:1–223 ISBN 978-952-10-5310-8. 55–63. https://doi.org/10.1016/J.FGB.2016.07.008. 52. Nakasone KK, Eslyn WE. A new species, Phlebia brevispora, a cause of 71. Brakhage AA. Regulation of fungal secondary metabolism. Nat Rev internal decay in utility poles. Mycologia. 1981;73:803–10. https://doi.org/10. Microbiol. 2013;11:21–32. 2307/3759792. 72. Martinez D, Berka RM, Henrissat B, Saloheimo M, Arvas M, Baker SE, et 53. Hood HM, Neafsey DE, Galagan J, Sachs MS. Evolutionary roles of upstream open al. Genome sequencing and analysis of the biomass-degrading fungus reading frames in mediating gene regulation in fungi. Annu Rev Microbiol. 2009; Trichoderma reesei (syn. Hypocrea jecorina). Nat Biotechnol. 2008;26:553– 63:385–409. https://doi.org/10.1146/annurev.micro.62.081307.162835. 60 https://www.nature.com/articles/nbt1403. Mäkinen et al. BMC Genomics (2019) 20:430 Page 22 of 22

73. Lo Presti L, Lanver D, Schweizer G, Tanaka S, Liang L, Tollot M, et al. Fungal 93. Sperschneider J, Catanzariti A-M, DeBoer K, Petre B, Gardiner DM, Singh KB, effectors and plant susceptibility. Annu Rev Plant Biol. 2015;66:513–45. et al. LOCALIZER: subcellular localization prediction of both plant and https://doi.org/10.1146/annurev-arplant-043014-114623. effector proteins in the plant cell. Sci Rep. 2017;7:44598. https://doi.org/10. 74. Raffaello T, Asiegbu FO. Small secreted proteins from the necrotrophic conifer 1038/srep44598. pathogen Heterobasidion annosum s.l. (HaSSPs) induce cell death in Nicotiana benthamiana. Sci Rep. 2017;7:8000. https://doi.org/10.1038/s41598-017-08010-0. ’ 75. Feldman D, Kowbel DJ, Glass NL, Yarden O, Hadar Y. A role for small Publisher sNote secreted proteins (SSPs) in a saprophytic fungal lifestyle: Ligninolytic Springer Nature remains neutral with regard to jurisdictional claims in enzyme regulation in Pleurotus ostreatus. Sci Rep. 2017;7:14553. https://doi. published maps and institutional affiliations. org/10.1038/s41598-017-15112-2. 76. Thuillier A, Chibani K, Belli G, Herrero E, Dumarçay S, Gérardin P, et al. Transcriptomic responses of Phanerochaete chrysosporium to oak acetonic extracts: focus on a new glutathione transferase. Appl Environ Microbiol. 2014;80:6316–27. https://doi.org/10.1128/AEM.02103-14. 77. Kajiwara S, Yamaoka K, Hori K, Miyazawa H, Saito T, Kanno T, et al. Isolation and sequence of a developmentally regulated putative novel gene, priA, from the basidiomycete Lentinus edodes. Gene. 1992;114:173–8. https://doi. org/10.1016/0378-1119(92)90571-6. 78. Bayry J, Aimanianda V, Guijarro JI, Sunde M, Latgé J-P. Hydrophobins—unique fungal proteins. PLoS Pathog. 2012;8:e1002700. https://doi.org/10.1371/journal.ppat.1002700. 79. Aimanianda V, Bayry J, Bozza S, Kniemeyer O, Perruccio K, Elluru SR, et al. Surface hydrophobin prevents immune recognition of airborne fungal spores. Nature. 2009;460:1117–21. https://doi.org/10.1038/nature08264. 80. Kim S, Ahn I-P, Rho H-S, Lee Y-H. MHP1,aMagnaporthe grisea hydrophobin gene, is required for fungal development and plant colonization. Mol Microbiol. 2005;57:1224–37. https://doi.org/10.1111/j.1365-2958.2005.04750.x. 81. Mali T, Kuuskeri J, Shah F, Lundell TK. Interactions affect hyphal growth and enzyme profiles in combinations of coniferous wood-decaying fungi of Agaricomycetes. PLoS One. 2017;12:e0185171. https://doi.org/10.1371/ journal.pone.0185171. 82. Salavirta H, Oksanen I, Kuuskeri J, Mäkelä M, Laine P, Paulin L, et al. Mitochondrial genome of Phlebia radiata is the second largest (156 kbp) among fungi and features signs of genome flexibility and recent recombination events. PLoS One. 2014;9:e97141. https://doi.org/10.1371/ journal.pone.0097141. 83. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15–21. https:// doi.org/10.1093/bioinformatics/bts635. 84. Hoff KJ, Lange S, Lomsadze A, Borodovsky M, Stanke M. BRAKER1: Unsupervised RNA-Seq-Based Genome Annotation with GeneMark-ET and AUGUSTUS. Bioinformatics. 2015;32:767–9. https://doi.org/10.1093/ bioinformatics/btv661. 85. Altschul SF, Wootton JC, Gertz EM, Agarwala R, Morgulis A, Schaffer AA, et al. Protein database searches using compositionally adjusted substitution matrices. FEBS J. 2005;272:5101–9. https://doi.org/10.1111/j.1742-4658.2005. 04945.x. 86. Koskinen P, Törönen P, Nokso-Koivisto J, Holm L. PANNZER: high- throughput functional annotation of uncharacterized proteins in an error- prone environment. Bioinformatics. 2015;31:1544–52. https://doi.org/10. 1093/bioinformatics/btu851. 87. Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics. 2005;21:3674–6. https://doi.org/10.1093/ bioinformatics/bti610. 88. Blin K, Wolf T, Chevrette MG, Lu X, Schwalen CJ, Kautsar SA, et al. antiSMASH 4.0—improvements in chemistry prediction and gene cluster boundary identification. Nucleic Acids Res. 2017;45:W36–41. https://doi.org/ 10.1093/nar/gkx319. 89. Petersen TN, Brunak S, von Heijne G, Nielsen H. SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods. 2011;8:785–6 https://www.nature.com/articles/nmeth.1701. 90. Emanuelsson O, Brunak S, Von Heijne G, Nielsen H. Locating proteins in the cell using TargetP, SignalP and related tools. Nat Protoc. 2007;2:953–71. 91. Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predicting transmembrane protein topology with a hidden markov model: application to complete genomes. J Mol Biol. 2001;305:567–80. https://doi.org/10.1006/ JMBI.2000.4315. 92. Sperschneider J, Dodds PN, Gardiner DM, Singh KB, Taylor JM. Improved prediction of fungal effector proteins from secretomes with EffectorP 2.0. Mol Plant Pathol. 2018. https://doi.org/10.1111/mpp.12682.